| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
Departments of
1
Pathology and Laboratory Medicine and
2
Internal Medicine, University Hospital Groningen, NL-9700 RB Groningen, The Netherlands
aaddress correspondence to this author at: Pathology and Laboratory Medicine, University Hospital Groningen, CMC-V, Room Y1.165, PO Box 30.001, NL-9700 RB Groningen, The Netherlands; fax 31-50-3612290, e-mail m.r.fokkema{at}path.azg.nl
Hyperhomocysteinemia, a risk factor for cardiovascular and thrombotic disease, pregnancy complications, and cognitive disorders, is defined as a fasting plasma total Hcy (tHcy) above a chosen cutoff value (1)(2). Reducing both the analytical and biological variation may add to the diagnostic value of any test. tHcy analytical variation (CVa) is method-dependent and ranges from 2.7% to 4.9% for fluorescence polarization immunoassays and from 2.5% to 14% for HPLC assays (3). The reported biological variations for tHcy under fasting or otherwise standardized conditions include 20–34% interindividual variation (CVg) and 7–11% (with a single extreme of 15–17%) intraindividual variation (CVi) (4)(5)(6)(7)(8)(9). Because of the short-term influence of meals, e.g., protein, on CVi (10), it is generally recommended that blood samples be collected under fasting or otherwise meal-standardized conditions. The influence of protein intake on between-day CVi is, however, unknown. The necessity for fasting has been questioned because fasting and nonfasting tHcy concentrations and reference values, as well as fasting and postprandial short-term CVi and CVg values, are similar (11)(12).
We investigated whether tHcy concentrations and variations under nonfasting conditions differ from those under fasting conditions. For this we calculated tHcy biological variation from samples taken at different clock times during a single day and during 3 weeks.
We studied 16 apparently healthy individuals [8 men and 8 women; median (range) age, 44 (25–58) years]. Our study design is illustrated in Fig. 1A
. Six blood samples were taken from all participants on Monday of week 1. A standardized breakfast (0815 in the morning) and lunch (1215) and a dinner (1730–1800) of unrestricted composition and quantity were provided during the day. Participants documented food and beverage intake on Monday and the preceding Sunday and fasted overnight from 2200 on Sunday. Participants were subsequently divided into two groups. Group 1 [five men and five women; age, 45 (25–58) years] took 5 mg of folic acid, 1 mg of cyanocobalamin, and 50 mg of pyridoxine daily at breakfast for 3 weeks (last intake on Sunday of week 4). On Monday of week 4, they repeated the protocol of Monday of week 1. Group 2 [three men and three women; age, 40 (25–57) years] had additional blood samples taken on eight occasions during weeks 1–3. They had no specific dietary restrictions apart from those on Monday of week 1. The study protocol was approved by the medical ethics committee of our hospital and was in agreement with local ethical standards and the Helsinki Declaration of 1975, as revised in 1996.
|
The compositions of the documented and standardized meals were calculated with the software package Becel (Hartog Union and Van den Bergh). Whole-blood vitamin B6 and vitamin B2 were determined by HPLC, serum folate and vitamin B12 by immunofluorometric methods (Autodelfia; Wallac Oy), and serum creatinine by a MEGA (Merck Darmstadt). tHcy was determined within 1 month of storage at -20 °C with a fluorescence polarization immunoassay (IMx; Abbott Laboratories) that measures tHcy, i.e., the sum of reduced and oxidized forms. tHcy samples from each individual were analyzed in a single run to minimize the contribution of CVa to total variation. The accuracy and precision of the tHcy method and the total, within-run, and between-run CVas were calculated according to NCCLS procedures (13), using calibrators (L-homocystine in human serum, from the IMx reagent set) at three concentrations and duplicate pool samples with a mean concentration of 15.3 µmol/L.
Between-day biological variation was calculated based on the nine underlined data points shown in Fig. 1A
for group 2. We did not use the ideal calculation method, i.e., nested ANOVA (14), because costs prevented us from measuring samples in duplicate. The between-day CVi (using the formula: SD x 100/mean tHcy) was calculated for each individual. After correction for within-run CVa [using the formula: CVi2 = CVi(measured)2 - CVa2], these calculations yielded a median, mean, and range for the between-day CVi. The median, mean, and range for between-day CVis for samples collected at 0800, 1200, and 1400 were calculated in a similar manner, using the three tHcy samples collected at those time points. The between-day CVg was calculated from the individual 3-week mean tHcy values, using their overall means and SD after correction for total CVa (tCVa). An individual’s mean provided an indication of that person’s homeostatic setpoint during the 3 weeks of the study. Between-day CVgs for samples collected at 0800, 1200, and 1400 were calculated from the individual means of the three tHcy samples collected at those time points, after correction for tCVa. Reliability coefficients [R = CVg2/(tCVa2 + CVi2 + CVg2)] and reference change values [RCV = Z x 
2 x (CVi2 + tCVa2), where Z = 1.96, i.e., the Z-value for the 95% confidence] were calculated. R values reflect the ability to estimate the average tHcy concentration for an individual (i.e., homeostatic setpoint) from a single tHcy measurement, whereas RCV values reflect the tHcy difference between two consecutive assays on different occasions that exceeds the combined tCVa and median CVi with 95% confidence (14). The within-day tHcy CVi, CVg, R, and RCV values were calculated from the six data points for groups 1 and 2 on Monday of week 1 and from the six data points for group 1 on Monday of week 4, i.e., after vitamin supplementation (for calculation, see the calculations for between-day biological variation).
Within-day tHcy changes on Monday of weeks 1 and 4 were investigated with repeated-measures ANOVA, followed by post hoc paired Student t-tests with Bonferroni corrections for 
, and day-mean tHcy changes were investigated with the paired Student t-test. CV and RCV differences were analyzed with F-tests and R differences with 
2 tests as described by Edwards (15). We used Spearman correlation tests to investigate correlations between within-day tHcy variation and changes on the one hand and vitamin concentrations, changes, and intake and protein intake on the other hand, as well as correlations between the between-day tHcy CVi and the 3-week mean vitamin concentration or variation.
Shown in Table 1
are the calculated analytical and biological variations for tHcy, together with between-day biological variations reported by others. Our method accuracy and precision were within limits, although one calibrator (at 7 µmol/L) showed a significant, but for this study irrelevant, difference of 0.32 µmol/L. Our "homeostatic" biological variation is comparable to values reported by others. Although nonsignificant, the 0800 samples seemed to have the highest between-day CVi, CVg, and RCV and the lowest R compared with samples taken at other clock times.
|
Shown in Fig. 1B
are the mean (95% reference interval) tHcy time courses on Monday of week 1 and Monday of week 4, i.e., before and after vitamin supplementation. There were significant within-subject changes (P = 0.004) before supplementation. tHcy decreased from 0800 to 1000 in the morning (P <0.0001), did not change between 1000 and 1400, and increased from 1000 to 1830 (P = 0.002) and from 1400 to 1830 (P = 0.002). Day-mean tHcy inversely correlated with day-mean whole-blood vitamin B2. Within-subject changes were also significant (P = 0.037) after supplementation: tHcy decreased from 0800 to 1000 (P = 0.024) and increased from 1000 to 1830 (P = 0.013). Day-mean tHcy decreased (P <0.0001), and although nonsignificant, within-day CVg and R seemed to decrease and RCV seemed to increase on vitamin supplementation. The within-day CVi did not change (Table 1
).
When data before and after supplementation were combined (n = 26), the day-mean tHcy inversely correlated with the day-means of serum folate, whole-blood vitamin B6, and whole-blood vitamin B2, but not with the day-means of serum vitamin B12. Relative tHcy changes from 0800 to 1200 (r = -0.406; P = 0.039) were significantly related to protein intake during the preceding evening. The between-day CVi inversely correlated with the 3-week mean vitamin B6 concentration.
The similarity between our biological variation coefficients and the standardized biological variations published by other investigators (Table 1
) suggests that it is not necessary to limit tHcy analyses to the fasting state. This is substantiated by the similarities of our time-specific CVi and CVg values. Fasting may also not be necessary for tHcy sampling during follow-up: our RCV was comparable (Table 1
) to RCVs reported in studies that collected samples in the fasting state or under otherwise standardized conditions (4)(5)(6)(7).
The within-day tHcy fluctuations observed in the present study, i.e., a decrease in tHcy after breakfast and an increase 2 h after lunch, are consistent with the findings of others (10)(16)(17). Our results confirm that CVg is particularly dependent on vitamin status (1)(18), based on the relationship between tHcy and vitamin status. The within-day CVi particularly depended on protein intake, as derived from the relationship between protein intake during the evening before sampling and the tHcy decrease from 0800 to 1200, which, together with the tHcy increase 2 h after lunch (i.e., at 1400), is consistent with amino acid uptake kinetics (10). Furthermore, our data suggest that both the CVg and the between-day CVi depend on protein intake to some extent because these seemed lowest at 1400 (i.e., after a 10-h fast and a subsequent 6-h period of low protein intake) and highest at 0800 (i.e., after a 10-h overnight fast). Sampling in the fasting state may thus be the least preferred, whereas long periods of low protein intake are likely to have the largest impact on reducing biological variation. Our study had insufficient power, however, to substantiate this difference and was also not designed to visualize the impact of protein intake standardization in detail. It remains questionable, however, whether a further reduction of biological variation is worth pursuing, notably when it requires the institution of patient-inconvenient protocols with probable poor compliance.
We conclude that a 10-h fast is not necessary and is probably a less-preferred metabolic condition for the standardization of diagnosis of hyperhomocysteinemia. Standardization of protein intake during the preceding evening might lower biological variation, but its influence is as yet unknown.
Acknowledgments
We thank Pim Modderman for assistance with the tHcy analyses. We gratefully acknowledge Herman Velvis and Enge Veenekamp-Hoolsemad for the vitamin B6 and vitamin B2 analyses. We also thank the staff of the immunochemical laboratory of our hospital for the folate and vitamin B12 analyses.
References
T mutation on homocysteine. Ann Clin Biochem 2000;37:390-398.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
